HR: 0800h
AN: H51B-0349    [Abstracts]
TI: Comparison of Evolutionary Multiobjective Algorithms For Calibrating An Integrated Semi-distributed Hydrologic Model
AU: * Tang, Y
EM: yxt132@psu.edu
AU: Reed, P
EM: preed@engr.psu.edu
AU: Wagner, T
EM: thorsten@psu.edu
AB: This study provides the first comprehensive assessment of state-of-the-art evolutionary multiobjective optimization (EMO) tools�_ relative effectiveness in calibrating integrated hydrologic models. The relative computational efficiency, accuracy, and ease-of-use of the following EMO algorithms are tested: Epsilon Dominance Nondominated Sorted Genetic Algorithm-II (ݏ-NSGAII), the Multiobjective Shuffled Complex Evolution Metropolis algorithm (MOSCEM-UA), and the Strength Pareto Evolutionary Algorithm 2 (SPEA2). This study assesses the performances of these three evolutionary multiobjective algorithms using a formal metrics-based methodology. This study uses two phases of testing to compare the algorithms�_ performances. In the first phase, this study uses a suite of standard computer science test problems to validate the algorithms�_ abilities to perform global search effectively, efficiently, and reliably. The second phase of testing compares the algorithms�_ performances for a computationally intensive multiobjective integrated hydrologic model calibration application for the Shale Hills watershed located within the Valley and Ridge province of the Susquehanna River Basin in north central Pennsylvania. The Shale Hills test case demonstrates the computational challenges posed by the paradigmatic shift in environmental and water resources simulation tools towards highly nonlinear physical models that seek to holistically simulate the water cycle. Specifically, the Shale Hills test case is an excellent test for the three EMO algorithms due to the large number of continuous decision variables, the increased computational demands posed by the simulating fully-coupled hydrologic processes, and the highly multimodal nature of the search space. A challenge and contribution of this work is the development of a comprehensive methodology for comprehensively comparing EMO algorithms that have different search operators and randomization techniques.
DE: 1805 Computational hydrology
DE: 1846 Model calibration (3333)
DE: 1847 Modeling
DE: 1873 Uncertainty assessment (3275)
SC: Hydrology [H]
MN: Fall Meeting 2005